Method and Apparatus for Intent Prediction and Proactive Service Offering

1. A computer implemented method for identifying customers across multiple channels, predicting customer intent, and proactively offering services to customers, comprising: a processor implemented automated speech self-service application framework using any of data that said framework persists, data persisted by other non-IVR interaction channels, which data persisted by non-IVR channels comprises any of recent Web page history and chat transcript contents, and enterprise data that said framework accesses, to greet customers and execute a personalized service strategy for each customer; said framework predicting the identity, or multiple possible candidate identities, of each customer; and said framework further comprising an intent prediction module configured to perform performance self-monitoring by regularly calculating a customer assessment success rate which comprises a percentage of times a customer assessment is completed in its entirety before conclusion of an assessment period, where the assessment period denotes a maximum amount of time available at the start of a customer interaction before prediction results are no longer valuable.

2. The method of claim 1 , further comprising: said framework generating a self optimizing confidence factor for each identity prediction that is adjustable after each interaction.

3. The method of claim 1 , further comprising: said framework predicting each customer's intent to estimate a most likely action or actions that said customer intends to take in an ongoing interaction; and said framework applying a self optimizing prediction confidence score to adjust said predicted customer intent on-the-fly and determine which assessment is most productively applied to the customer's interaction.

4. The method of claim 1 , further comprising: said framework proactively offering to the customer any of at least one solicitation for service that is relevant to said customer's current interaction and a desired business outcome.

5. The method of claim 1 , further comprising: said framework calculating a customer tolerance score that models how tolerant the customer is likely to be with regard to the use of IVR automation in general or in connection with a specific type of customer interaction.

6. A system for predicting customer intent in connection with an IVR session, comprising: a customer care center in communication with a customer via a multichannel communication network; said customer care center configured for responding to a customer interaction and, upon receiving confirmation of said customer's identity, predicting the intent of said customer's interaction; wherein said customer comprises one among many customers who interact with said customer care center, each of which has a specific intention to receive a desired service or information; and said customer care center further comprising an intent prediction module configured to perform performance self-monitoring by regularly calculating a customer assessment success rate which comprises a percentage of times a customer assessment is completed in its entirety before conclusion of an assessment period, where the assessment period denotes a maximum amount of time available at the start of a customer interaction before prediction results are no longer valuable.

7. The system of claim 6 , said customer care center comprising any one or more of a combination of multiple customer support mechanisms which comprise any of an IVR system, one or more customer care executives, a Web site, an email support system, an SMS support system, a chat-based support system, one or more native mobile applications for a phone or tablet, and one or more offline support systems.

8. The system of claim 6 , said multichannel communication network comprising any one or more of a plurality of communication networks which comprise any of SMS, Web, email, mobile and outbound networks.

9. The system of claim 6 , said customer care center further comprising: a customer service support system comprising an expert service portal (ESP) block, a real-time caller assessment framework, a customer interaction database for augmenting customer data repositories, interaction capture services, and a prediction model refinement mechanism each of which, alone or in combination, identifies said customer, predicts customer intention, calculates propensities for specific customer behaviors to be successfully elicited, and minimizes duration of customer interaction with said system.

10. The system of claim 6 , said customer care center further comprising: a customer entry point, in communication with said customer care center, for interacting with customers through said multichannel communication network, wherein said customer is associated with any of different languages, regions, products, and departments.

11. The system of claim 6 , said customer care center further comprising: in response to interaction by said customer with a customer entry point, an expert service portal (ESP) comprising a greeting module for welcoming said customer and simultaneously identifying said customer based upon an incoming phone number when customer interaction is over phone or by SMS, based upon a username when the interaction is on the Web or in a mobile application, and based upon an email ID when customer interaction is via email or offline support.

12. The system of claim 11 , said greeting module further comprising: a mechanism that determines a confidence score based on multiple factors comprising any of a history of previous access by said customer, previous success rate in identifying said customer, customer care executive input from an agent who interacted with a previous communication from a same phone number or email ID, an explicit opt-in action by said customer that identifies the customer every time that the customer calls from the same number, and call history from the same phone number.

13. The system of claim 6 , said customer care center further comprising: an identity confidence scoring mechanism for developing and managing a storage mechanism in a customer interaction database that contains unrestricted, many-to-many relationships between a phone number and customer IDs, said identity confidence scoring mechanism correlating a name with an ID to evolve on a continuous or discrete scale, based on inputs observed by the system.

14. The system of claim 13 , said identity confidence scoring mechanism further comprising: a mechanism that addresses phone numbers that are shared by multiple customers by using parallel scoring and assessment of multiple customer IDs for an interaction, and by selectively applying either an interactive disambiguation process or a sense of a personalized interaction with multiple identities under approved conditions.

15. The system of claim 6 , said customer care center further comprising: when an identity confidence score is low and ambiguity exists about a customer's identity, an identity module that performs an identity check by asking said customer about the customer's identity directly; and when an identity confidence score is high, said identify module guessing a customer's identity using an identity prediction module and confirming said customer's identity based upon said customer's response to said guess.

16. The system of claim 15 , further comprising: said identify module configured to use an outcome of a current guess to determine a confidence score in future interactions with customers having the same identity.

17. The system of claim 6 , said customer care center further comprising: after a customer's identity is confirmed, an intent prediction module configured to use multiple data sources to contribute to prediction of customer intent for a current interaction.

18. The system of claim 17 , said multiple data sources comprising any of: recent transactions of calls received from a same number; recent SMS communications from a same number; recent email communications with the same customer; recent web browsing and mobile devices applications sessions; trends and patterns of all inbound and outbound phone communications with the likely identity of the customer; and use of information recorded about interactions in non-phone channels, including any of a call center agent, web site, mobile application, store, and direct mail.

19. The system of claim 6 , said customer care center further comprising: an intent prediction module for predicting said customer intent; and a prediction validation module for validating predictions from said intent prediction module.

20. The system of claim 19 , further comprising: said intent prediction module configured to predict any of whether the customer should be provided with accelerated service when an intent prediction is validated to a predetermined level of confidence, and whether the customer should be provided with personalized service when an intent prediction is not validated to a predetermined level of confidence.

21. The system of claim 6 , said customer care center further comprising: in response to said prediction, a mechanism configured for determining an optimal strategy to apply to said customer interaction via any of: a contact center infrastructure that invokes a self service module or live service for sales or services; customer interaction language; a product or service that a customer is looking for or recently bought; a type of customer; a best style to address the customer; and next steps in a business process.

22. The system of claim 6 , said customer care center further comprising: a prediction model refinement module configured to implement a closed-loop prediction optimization algorithm for learning a combination of factors which most accurately predict actual activity during a customer interaction.

23. The system of claim 6 , said customer care center further comprising: a prediction module configured for using an intent confidence score for intent prediction and, depending on a confidence score and intent of the customer, for creating a service strategy.

24. The system of claim 23 , wherein said service strategy comprises any of: directing the customer to a self help system; sending an SMS to the customer's mobile number with necessary information that the customer is looking for; sending an email to the customer with information that customer is looking for; passing the customer to a customer care executive; directing the customer to chat support; asking the customer to go to a self help portal; and scheduling a call back to the customer.

25. The system of claim 6 , said customer care center further comprising: a module configured for calculating caller tolerance to the use of IVR automation on an interaction, based on computing scores for prior customer effort levels to use an IVR system and recent success rates in using an IVR system.

26. The system of claim 6 , further comprising: a customer preferences dialog comprising any of: permanent opt-in to, and/or opt-out of a customer assessment and recognition process; temporary exclusion on a per interaction basis from the customer assessment and recognition process; and one or more proffered spoken languages or language variants to be used on future interactions.

27. The system of claim 6 , further comprising: a single cross-enterprise mechanism managed centrally via a web portal and configured to provide a customer experience that varies according to any of: playing of branding messages; announcement of delivery; primary interaction language; secondary language options for callers; menu options; association with a line of products or services; and identifying a task that the customer is expected to perform during a customer interaction.

28. The system of claim 25 , further comprising: computing said scores based upon customer effort levels and success rates earlier in a current call.